Hominin Evolution

Scientific ocean drilling can transform our understanding of the role of past African climate changes on early human evolution. Current hypotheses of human origins suggest that late Neogene changes in northeast African climate influenced the evolution of human ancestral lineages, leading to the emergence of traits that are uniquely human: bipedality, exceptionally large brains, and the construction of increasingly sophisticated stone tools.

The basic premise of these hypotheses is that large-scale shifts in climate alter the ecological composition of a landscape which, in turn, presents specific faunal adaptation or speciation pressures leading to genetic selection and innovation. Emerging fossil faunal and paleoclimate data broadly support this idea but fundamental questions remain concerning the timing, nature, and causes of African climate variability and for defining the imprint of climate change on the fossil record of faunal evolution. ―The field of paleoanthropology is on the brink of novel ideas and datasets concerning the how and why of hominin evolution. Tremendous advances in the environmental sciences are forcing these developments to occur‖ (Potts, 1998).

The rationale supporting a major drilling initiative to understand the role of past climate change in human evolution is compelling. There are presently very few ocean drilling sites off East Africa, where the majority of hominin fossils are found, and none of these are suitable for comprehensive paleoclimate research. While some DSDP sites drilled over thirty years ago in the Gulf of Aden have been useful for low-resolution paleoclimate records (Feakins, 2006), these were rotary drilled and have significant

recovery gaps and disturbed intervals. No shallow (carbonate-bearing) sites exist along the entire east coast of Africa, nor are any sites available near the mouths of major river systems draining the East African interior. Most of our current understanding of East African paleoclimate changes comes from marine records of aeolian dust export from sites off subtropical west and northeast Africa (deMenocal, 2004) and it can be reasonably argued that these records are distal from the fossil localities themselves.

Ocean drilling has a unique opportunity to contribute observations to test the hypothesized role of past climate changes in shaping the course of human evolution.

Terrestrial paleoclimate records from East African fossil localities or from lake basin drilling programs are either too short or stratigraphically incomplete. Ocean drilling of specific, previously-undrilled sedimentary packages described below provides the best opportunities for constraining the timing, nature, and causes of paleoclimatic change in this geographic region where our ancestors evolved.

6.2.1 Key events and environmental hypotheses of human origins

Bursts of evolutionary activity and behavioral changes punctuate the fossil record of early human evolution (Fig. 6.9). Although the fossil record is still incomplete, two time windows encompass key evolutionary events near 3.0–2.5 Ma and 2.0–1.5 Ma that effectively shaped the characteristics that define us as human (see summary in Potts, 2006). Impressively, these same time windows also include dramatic changes in other African mammalian taxa such as bovids and rodents (Vrba, 1995).

Between 3.0–2.5 Ma at least two new hominin lineages emerged (Paranthropus and early Homo genera) from an ancestral lineage that itself became extinct at this time (Australopithecus afarensis); this interval also marks the first appearance of stone tools (Semaw et al., 2003) (Fig. 6.9). Members of the Paranthropus lineage are distinctive for their robust, massive frames, large and broad post-canine (molar) dentition, specialized chewing adaptations (sagittal crest), and intermediate cranial volumes. The Paranthropus lineage first appeared at ~2.7 Ma (Fig. 6.9). The fossil record of African bovids (antelope family) indicates evidence for exceptional faunal first and last appearances near this time, roughly synchronous with the appearance of specialized arid-adapted grazers (Vrba, 1995; Bobe and Behrensmeyer, 2004; Bobe et al., 2002).

The earliest record of stone tools, at ~2.6 Ma, is from Gona, Ethiopia (Oldowan industry) (Semaw et al., 2003).

By 1.8–1.6 Ma, Homo habilis became extinct and its immediate successor and our more direct ancestor, H. erectus, first occurs in the fossil record near 1.8 Ma (Kimbel, 1995). Homo erectus may have migrated to southeast Asia as early as 1.9–1.8 Ma (Swisher et al., 1994). Near 1.7 Ma, South African (Reed, 1997) and East African bovid assemblages shifted toward further absolute increases in the abundance of arid-adapted species (Vrba, 1995). Earliest occurrences of the more sophisticated Acheulean tool kit (bifacial blades and hand axes) occurred near 1.7–1.6 Ma (Ambrose, 2001; Clark et al., 1994) (Fig. 6.9).

Environmental hypotheses of early human evolution share the view that changing African environmental conditions over the late Neogene selected for the morphological and behavioral characteristics that make us human. Where hypotheses differ is in the proposed role of climate change in natural selection. The Savannah Hypothesis is perhaps the best known of the habitat-specific hypotheses of African faunal evolution

(Bartholomew and Birdsell, 1953; Dart, 1925; Klein, 1989; Wolpoff, 1980). Current interpretations of the Savannah Hypothesis state that the evolution of African mammalian fauna, including early hominins, was primarily linked to the progressive expansion of more open grassland conditions. The Turnover Pulse Hypothesis is a recent variant of this idea and posits that focused bursts of biotic change (quantified in terms of first and/or last appearance datum clustering) were initiated by progressive shifts toward greater African aridity that occurred roughly near 2.8 Ma and 1.8 Ma (Vrba, 1985, 1995). The Variability Selection hypothesis accommodates one of the more obvious yet also curious features of the fossil record (Potts, 1998). Fossil hominin and other mammal lineages typically persisted over long durations (10⁵ to 10⁶ year) yet they are preserved within sediment sequences recording rapid, orbital scale (10³ to 10⁴ year) climate oscillations (deMenocal, 1995; Dupont and Leroy, 1995; Feibel et al., 1989). This view suggests that changes in the amplitudes of orbital African climate variability, linked to the eccentricity modulation of precessional monsoonal cycles, may have been an important genetic selection criterion. A fundamental limitation to testing these hypotheses has been the lack of suitably detailed, well-dated, and multi-proxy reconstructions of African climate variability to constrain the basic paleoclimate history of the region. How and why did African climate change?

Figure 6.9 Summary figure from Feakins et al. (2005). A) Interval means and (1) standard deviation of C30 n-alkanoic acid ¹³C for intervals >40 k.y. in duration (filled circles) and for intervals <40 k.y.

induration (open circles, dashed lines). B) Soil carbonate ¹³C from Turkana Basin (northern Kenya);

means and (1) standard deviations from individual stratigraphic layers (from sources in Feakins et al., 2005). C) Phylogeny of major hominin lineages throughout Pliocene-Pleistocene (from sources in deMenocal, 2004).

6.2.2 African paleoclimate history

Over the late Neogene, North Africa and equatorial East Africa experienced a long-term drying trend that commenced near 3 Ma, culminating in the driest, most open conditions around 1.8 Ma. Superimposed upon this aridification trend are rapid, orbitally paced wet-dry cycles that persisted throughout the Neogene (see summary in deMenocal, 2004).

Evidence documenting this pronounced shift to more open conditions has been found in carbon isotope analyses of East African soil carbonate nodules, indicating a dramatic increase in the proportion of C4 (grassland) vegetation that commenced near 3 Ma and reach maximum development near 1.8 Ma (Fig. 6.9) (Cerling, 1992; Cerling et al., 1994;

Wynn, 2004). At Gulf of Aden DSDP Site 231, a comparable record of C4 grassland expansion was established using carbon isotope analyses of plant wax biomarker compounds (Feakins, 2006) (Fig. 6.9). At Site 231 there is a 40% increase in C4 representation over the last 4 Ma from a nearly pure C3 baseline in the early Pliocene.

Higher resolution analyses at Site 231 also resolved orbital-scale (precessional) vegetation cycles and, surprisingly, these were found to have peak-to-peak amplitudes that were as large as the overall late Neogene trend.

Figure 6.10 Evidence for high- and low-amplitude African climate variability ‗packets‘ from West and East African Sites 659 and 721/722 (deMenocal, 1995; Tiedemann et al., 1994). African paleoclimate variability over this interval was strongly regulated by orbital precession, which is modulated by orbital eccentricity (shown). Prolonged (10⁴-10⁵ year) intervals of exceptionally high- or low-amplitude paleoclimate variability are apparent off both margins of subtropical Africa. Highest variability occurs during periods of maximum orbital eccentricity (e) when modulation of the precession index (Π = e*sin (ω)) and the seasonality of low-latitude insolation receipt is greatest (Berger, 1978; Prell and Kutzbach, 1987). Low variability intervals are shaded. The filled black rectangles denote lactustrine depositional phases between 2.51–2.66 Ma recorded at several NE African terrestrial localities (Trauth et al., 2005).

Figure modified from deMenocal, 2004.

African paleoclimate history also includes persistent orbitally paced wet-dry climate cycles throughout the Neogene linked to orbital precession regulation of African monsoonal rainfall intensity. The Mediterranean sapropel record is one of the most compelling examples of orbital climate control. Ocean drilling in the eastern Mediterranean and careful fieldwork on older, uplifted sediments in southern Italy have documented the strong eccentricity and precessional response of sapropel deposition related to episodes of enhanced Nile River outflow linked to orbital monsoon forcing.

These same precessional climate cycles are also detected in aeolian dust records from ODP sites off East and West Africa (Clemens et al., 1996; deMenocal, 1995; Tiedemann et al., 1994). These eccentricity-modulated variability ‗packets‘ represent 10⁴-10⁵-year intervals of exceptionally high- or low-amplitude paleoclimatic variability. Some of the largest amplitude packets are associated with widespread deposition of deep lake diatomite facies in East African Rift basins (Trauth et al., 2005) (Fig. 6.10).

Ocean drilling can significantly contribute to dating the fossil record of human evolution. Dates for most fossil material are obtained by either direct radiometric dating

Figure 6.11 Volcanic ash shard abundances at DSDP Site 231 (modified from Feakins et al., 2007).

Areas of known volcanic activity during the Pliocene and Pleistocene are identified with asterisks.

Approximately 15-20 tephra layers are identified for the Pliocene-Pleistocene interval. Continuous scrape-sampling provides a complete downcore sedimentary record throughout the last 4Ma where sediments were available (solid); gaps in the sediment recovery indicate areas where no data on tephrostratigraphy are available (recovery was about 70%), so many tephra layers were probably not sampled.

or indirect stratigraphic dating of fossil material intercalated between datable volcanic tuff deposits. Since the East African Rift region has been volcanically active for much of the late Neogene, the rift basins have been periodically blanketed with tephra from hundreds of discrete volcanic eruptions. A relatively small number of these tuffs have sufficiently large/abundant phenocrysts for direct (^40/30Ar) dating and the majority of ash layers serve as marker beds with approximate (interpolated) ages. These same ash layers are found and indentified geochemically in marine sediments offshore of East Africa (Fig. 6.11). Here they can be dated directly using the orbitally tuned oxygen isotope stratigraphy. This has been demonstrated for some ash layers preserved in Arabian Sea and Gulf of Aden sediments (deMenocal and Brown, 1999; Feakins et al., 2007), but a larger, all-Africa integration of tephra stratigraphy with the marine oxygen isotope stratigraphy awaits new core material.

6.2.3 East African river systems as archives of regional climate variability

To constrain African paleoenvironmental changes during the late Neogene will require the recovery of sediments that specifically address the timing and signatures of African climate change over the subtropical African geographic domain where hominin fossils are found, including South Africa, Tanzania, Kenya, and Ethiopia. Ocean drilling targets would need to be prioritized by their anticipated scientific contribution to the primary question of whether climate change impacted African faunal evolution. This section addresses the most valuable drilling targets – sites that, if drilled, would reshape our understanding of the timing and causes of African climate changes over the period of major African faunal evolutionary changes.

Jubba, Rufiji, Limpopo, Orange, and Zambezi distal fan drilling

Among the most promising new opportunities for reconstructing past changes in African climate are sediment packages accumulating near the mouths of rivers draining large sectors of subtropical and tropical Africa (Fig. 6.12). Hemipelagic sediments accumulating near the mouths of rivers draining interior East and South Africa (Jubba, Rufiji, Limpopo, Zambezi, and Orange Rivers) would provide promising opportunities for understanding both terrestrial paleoclimate and regional paleoceanographic variability over the Neogene. Smaller river systems in Kenya (Galana and Tana) and Tanzania (Rufiji) were likely much larger and more active systems in the past when African monsoonal rainfall was stronger, based on lithologic evidence of much larger and deeper East African lake systems (Garcin et al., 2009; Trauth et al., 2005).

The largest of these East African drainage basins are, from north to south (Fig.

6.12), the Jubba (800,000 km²), Rufiji (175,000 km²), Limpopo (415,000 km²), and Zambezi Rivers (1,400,000 km²). The Jubba drainage basin spans roughly one-third of the national area of Kenya, Somalia, and Ethiopia and captures summer monsoonal rainfall, principally off the Ethiopian highlands. The Rufiji drainage basin is the smallest of the three and is confined to eastern Tanzania. The Zambezi drainage basin is the largest of the three (the fourth largest river in Africa) and the drainage basin includes much of Zambia, Malawi, and Zimbabwe. Significantly, the Zambezi drainage basin also includes Lake Malawi, so lacustrine sequences drilled there may be directly compared to offshore drilled sequences of the Zambezi distal fan.

Figure 6.12 African drainage basins (modified from Africa Earth Observatory Network Database http://www.aeon.uct.ac.za).

Africa‘s largest drainage basins, such as the Nile, Niger, and Congo, each drain several millions of square kilometers and thus represent large areal integrators of regional climate change. Smaller drainage basins of East Africa, such as the Ganane, Rufiji, Zambezi, and Orange rivers, drain terrains containing known hominin fossil localities.

The sediments accumulating in the modest deltaic systems formed by these rivers are comprised of proximal and distal fan deposits. Proximal fan sediments are commonly complicated by turbidites and intermittent sedimentation, whereas distal fans provide more continuous, high accumulation rate, hemipelagic depositional environments. Distal fan sediments are comprised of marine biogenic components (microfossils and marine organic carbon) as well as terrestrial lithogenic (riverine clays and silts) and organic (terrestrial organic matter and biomarkers) sediments. Oxygen isotope analyses of marine foraminifera can be used to constrain an orbital (10⁴ year) chronology. The terrestrial organic fraction can be exploited to yield an impressive diversity of proxies that monitor the paleoclimatic, paleohydrological, and paleovegetational history of the specific drainage basin.

Analyses of the terrestrial organic fractions in distal fan environments have presented new opportunities for reconstructing African vegetation and hydrologic changes. River-borne terrestrial organic matter is comprised of a broad spectrum of compounds, or biomarkers, which can be used to record basin-scale changes in regional climate, vegetation, and hydrological balance. For example, a recent study of a sediment core off the Congo fan documented large changes in Congo Basin mean annual temperature using terrestrial biomarkers that record river runoff (Branched and Isoprenoid Tetraether (BIT) index) and soil temperatures (Methylation of Branched Tetraethers (MBT) index), as well as the D composition of plant wax biomarker to document changes in regional humidity changes (Weijers et al., 2007). Another related study from this same area used ¹³C analyses of plant wax biomarkers (n-alkanes) to document glacial-interglacial changes in the relative proportion of C3-C4 vegetation in the Congo Basin. These are examples of the approaches that could be applied to East African nearshore sediments recovered under this drilling mission.

Gulf of Aden drilling

The Gulf of Aden is the most proximal ocean basin to hominin fossil localities in Ethiopia, Kenya, and Tanzania and several previous studies on older DSDP material have demonstrated the utility of these sediments for addressing this problem. Several cruises lead by French, Japanese, and American scientists have acquired single- and multichannel seismic lines for the region to identify drill sites. Piston cores obtained at these drill sites indicate high accumulation rates (5-10 cm k.y.^-1) at these localities.

Unfortunately drilling in this region is likely to be challenging given the geopolitical tensions of the Gulf of Aden region.

Together, drilling these regions (the Gulf of Aden and the distal fan deposits from the Jubba, Rufiji, and Zambezi Rivers) would fundamentally reshape our understanding of the timing and causes of African climate changes over the period of major African faunal evolutionary changes.